polymers

Article
Recycling of Bottle Grade PET: Influence of HDPE
Contamination on the Microstructure and Mechanical
Performance of 3D Printed Parts
Joanne Vaucher, Adrien Demongeot , Véronique Michaud and Yves Leterrier *

Laboratory for Processing of Advanced Composites (LPAC), École Polytechnique Fédérale de Lausanne (EPFL),
CH-1015 Lausanne, Switzerland
* Correspondence: [email protected]

Abstract: As part of a project that aims to provide people with disabilities with simple assistive
devices in Colombia, the possibility of creating a PET filament that can be printed by Fused Deposition
Modelling (FDM) from beverage bottle waste was investigated, with the aim to remain as simple as
possible in terms of plastic collection, sorting, processing, and printing. Recycled PET filaments were
thus produced by extrusion from collected PET bottles, with the potential addition of HDPE, which
comes from caps and rings. The microstructure, mechanical performance, and printing quality of
parts produced with these filaments were investigated in comparison to commercial PET virgin and
recycled filaments. HDPE presence as an immiscible blend did not affect the ease of extrusion or the
quality of the printing, which were all satisfactory. In some conditions, the addition of 5 wt% of HDPE
to recycled PET had a toughening effect on otherwise brittle samples. This behavior was attributed
to the presence of elongated HDPE inclusions resulting from shear forces induced by the layer-by-
layer printing, provided that the interface temperature remained high between layer depositions.
Citation: Vaucher, J.; Demongeot, A.; This confirms that the mechanical performance of recycled PET is very sensitive to the processing
Michaud, V.; Leterrier, Y. Recycling of conditions, especially in the case of 3D printing. Nonetheless, this low-cost process that did not
Bottle Grade PET: Influence of HDPE require sophisticated compatibilization schemes allowed for the printing of parts with mechanical
Contamination on the Microstructure properties comparable to those obtained with high purity, commercially recycled filaments, opening
and Mechanical Performance of 3D interesting perspectives for a low-cost PET recycling process.
Printed Parts. Polymers 2022, 14, 5507.
https://doi.org/10.3390/ Keywords: recycling; contamination; poly(ethylene terephthalate); high density polyethylene;
polym14245507 3D printing
Academic Editors: Mario Bragaglia
and Francesca Nanni

Received: 14 October 2022

1. Introduction
Accepted: 12 December 2022
Published: 15 December 2022
In the last decade, the production of polymer-based materials has significantly in-
creased. In 2019, the global plastic production reached 370 million tons, and this amount is
Publisher’s Note: MDPI stays neutral
expected to increase to 900 million tons by 2050 [1,2]. Packaging represents the dominant
with regard to jurisdictional claims in
application market, accounting for more than 40% of the global consumption of plastics [3].
published maps and institutional affil-
Polyethylene terephthalate (PET) is one of the most favorable packaging materials, thanks
iations.
to its light weight, clarity, and barrier properties against moisture and oxygen [4]. PET
packaging represents almost half of single-serve beverage packaging and 12% of global
solid waste [5]. The consequences of the worldwide use of plastic on the environment are
Copyright: © 2022 by the authors.
important, especially in landfills and oceans, and the urge to develop efficient recycling
Licensee MDPI, Basel, Switzerland. solutions is central to the scientific community. Recycling of PET is already a widely devel-
This article is an open access article oped field worldwide, and various routes, including mechanical, chemical, and biocatalytic
distributed under the terms and recycling, are investigated [6,7]. In Switzerland, more than 80% of the consumed PET
conditions of the Creative Commons bottles are collected and recycled, mainly into new beverage bottles [8,9]. This closed-loop
Attribution (CC BY) license (https:// recycling process was optimized in terms of material purity to avoid contamination issues
creativecommons.org/licenses/by/ with foreign materials and molecular weight control to guarantee stable processing and per-
4.0/). formance. Indeed, contamination of PET with other types of polymers, such as high-density

Polymers 2022, 14, 5507. https://doi.org/10.3390/polym14245507 https://www.mdpi.com/journal/polymers

Polymers 2022, 14, 5507 2 of 15

polyethylene (HDPE), coming, for instance, from the bottles’ caps, affects the mechanical
properties of injection-molded recycled PET due to phase separation in the immiscible
blend [10]. The compatibilization of these two polymers using reactive modifiers such as
maleic anhydride is well-established but requires accurate processes [11–13].
A novel method for processing virgin and recycled PET is fused deposition modeling
(FDM) [14,15], one of the leading additive manufacturing processes in which layers of an
extruded filament of the polymer are applied sequentially on a carrier into a predefined
pattern, with the potential to create any 3D architecture. The use of PET in 3D printing is not
widespread nowadays, primarily because of its semi-crystalline nature, and the influence
of contamination on the quality of printed parts has not been explored. Indeed, semi-
crystalline polymers often negatively impact the printing process, for instance in cases of
excessive shrinkage [16] and may result in brittle parts. In fact, polyethylene terephthalate
glycol-modified (PETG) is usually preferred in 3D printing, as the substitution of ethylene
glycol by cyclohexanedimethanol in the PET backbone makes it more flexible and less
brittle when printed [17–19].
As part of a project that aims to provide people with disabilities with simple assistive
devices in Colombia, the possibility of creating a printable PET filament from beverage
bottle waste was investigated, since the availability of commercial filaments is limited in
many remote areas. The novelty was to check whether the contamination of the recycled
PET with HDPE would be critical for the mechanical performance of printed parts in the
absence of compatibilization. Indeed, the process had to be as simple as possible due to
the limited infrastructure available, both in terms of plastic collection, sorting, processing,
and printing. Thus, this study focuses on the low-cost recycling of bottle-grade PET into
filaments for 3D printing by FDM, with a focus on the effect of controlled concentrations
of HDPE on the structure and properties of 3D printed parts. The interplay between
printing conditions, blend microstructure, and resulting mechanical properties is analyzed
to propose optimal compositions and printing conditions. This investigation shows that it is
possible to 3D-print HDPE-contaminated PET parts with excellent mechanical performance,
but only under specific process conditions.

2. Materials and Methods

2.1. Materials and Processes
PET and HDPE were collected from used 500 mL bottles at the EPFL in Lausanne,
Switzerland. All labels were torn off by hand, and surfaces with remaining traces of glue
were cut out. The HDPE cap and ring closure were carefully separated from the PET bottles,
and both types of plastic were washed with cold water and soap, then dried in the air
at room temperature for at least 24 h. The two polymers were ground using a shredder
(150 Series, Rapid Granulator, Sweden) with a 5 mm-diameter hole perforated screen.
No additional drying was performed on the materials prior to subsequent processing
and testing. This was a deliberate choice due to the limited infrastructure available for
disabled people in Columbia, which could impact the final performance of the printed
materials [20]. The ground polymer flakes were mixed with a spoon in batches of 50 g to
achieve controlled concentrations of HDPE in PET ranging from 0 to 10 wt%. The polymer
mixtures were extruded into filaments with a compounding twin screw extruder (Prism
TSE 16 TC, Thermo Fisher Scientific) at temperatures of 225 ◦ C at the feeder, 230 ◦ C in the
center, and 235 ◦ C at the die, with a screw velocity of 30 rpm. Cooling of the filaments
was carried out on a rolling bed at ambient temperature with the help of two domestic
fans. The diameter of the die was 2 mm, and the targeted filament diameter was 1.75 mm.
A summary of the various types of filaments is given in Table 1. In order to facilitate the
identification of the contamination level, HDPE batches were sorted by color, and each
blend was prepared with a different color. In preliminary tests, it was found that the effect
of the color pigment in the HDPE phase was negligible, as could be expected due to the
immiscible nature of HDPE, combined with its rather low concentration. For comparison
Polymers 2022, 14, 5507 3 of 15

purposes, commercial filaments of virgin PET (EPR InnoPET, Innofil) and recycled PET
(rPET Natural Blue) were purchased from BASF 3D Printing Solutions, The Netherlands.

Table 1. Characteristics of the extruded filaments.

PET Fraction HDPE Fraction Diameter

Nomenclature Color
[wt%] [wt%] [mm]
E-PET0 100% 0% Transparent 1.72 ± 0.03
E-PET2 98% 2% Grey 1.75 ± 0.04
E-PET5 95% 5% Yellow 1.76 ± 0.03
E-PET10 90% 10% Orange 1.71 ± 0.06

Samples were printed from commercial and home-extruded filaments with a Prusa i3
MK3 printer (Czech Republic) at die and bed temperatures of 265 ◦ C and 100 ◦ C, respec-
tively. A 0.4 mm diameter die and a standard polyetherimide bed, wiped with a commercial
cleaning agent, were used. A homemade insulating box with a metallic frame on which
40 mm-thick extruded polystyrene boards were assembled to cover each face was placed
around the printer in order to control the surrounding temperature. The printing quality
was investigated through the analysis of 3D benchy models. Those objects representing a
tugboat were designed on purpose to test the printer’s accuracy and contained most of the
difficult-to-print features, such as symmetry, overhanging curved surfaces, smooth surfaces,
holes, etc. [21]. Each 3D benchy model was printed with a layer height of 0.2 mm, 10%
infill, an extrusion print speed of 50 mm/s, and a travel speed of 120 mm/s. Specimens for
tensile testing were printed in the 0◦ direction (parallel to tensile testing) with an infill of
100% and a randomized infill start for all types of filaments. Measurements were carried
out on 3 different specimens for each batch. Two series were prepared: in the first one,
specimens were printed one by one, with 27 s of cooling time between the deposition of
consecutive layers, while in the second one, 3 specimens were printed simultaneously each
time, with 85 s of cooling time between consecutive layers. These two series are respectively
termed ‘single prints’ and ‘3 prints’ in the following.
Injection-molded specimens were also prepared with cuts of E-PET0 and E-PET5
filaments using a Micro Injection Moulders Xplore IM 5.5 (the Netherlands), with a melt
temperature of 270 ◦ C, a mold temperature of 80 ◦ C, and a pressure of 8 bar maintained
for 7 s.

2.2. Characterization Methods

The polymer samples were characterized using different methods, as summarized in
Table 2. Thermogravimetric analysis (TGA) of polymer flakes was performed on a Perking
Elmer TGA 4000 with a sensitivity of 1 µg. The samples were heated from 30 to 600 ◦ C at a
rate of 10 ◦ C/min under a 20 mL/min air flow to mimic the extrusion environment. A sec-
ond analysis was performed according to the following heating procedure: heat from 30 to
265 ◦ C with a rate of 10 ◦ C/min, then hold at this temperature for 30 min. This temperature
corresponds to the printing temperature of the filament, and the experimentation was done
in order to highlight any degradation that would not occur instantaneously. The samples
were not dried before the measurements in order to replicate the real processing conditions.
Tensile tests were performed according to the ISO 527-1 standard [22] on type 5A
specimens (gauge length and width of 25 mm and 6 mm, respectively) with a ZWICK 10 kN
uniaxial testing machine (Germany), using clip-on extensometers. Measurements were
carried out on 3 different specimens for each batch, along the 0◦ printing direction. Tensile
strength, elongation at break, and Young’s modulus were determined as recommended by
the ISO 527-1 standard [22].
Polymers 2022, 14, 5507 4 of 15

Table 2. Summary of preparations and tests.

Name Composition TGA DSC Tensile Tests SEM

PET Flakes of recycled PET x
HDPE Flakes of recycled HDPE x
E-PET0 Extruded recycled PET x x
E-PET2 Extruded recycled PET with 2 wt% recycled HDPE x x
E-PET5 Extruded recycled PET with 5 wt% recycled HDPE x x
E-PET10 Extruded recycled PET with 10 wt% recycled HDPE x x
P-PET0 Printed recycled PET x x x
P-PET2 Printed recycled PET with 2 wt% recycled HDPE x x x
P-PET5 Printed recycled PET with 5 wt% recycled HDPE x x x
P-PET10 Printed recycled PET with 10 wt% recycled HDPE x x x
I-PET0 Injected extruded recycled PET (after extrusion) x x x
I-PET5 Injected extruded recycled PET with 5 wt% recycled HDPE x x x
PET_C Commercial virgin PET filament x
PET_CR Commercial recycled PET filament x
P-PET_C Printed commercial virgin PET x
P-PET_CR Printed commercial recycled PET x

Differential scanning calorimetry (DSC) measurements were performed on 10 mg

samples using a DSC Q100 from TA Instruments (USA) following a first heating cycle from
25 to 300 ◦ C at 10 ◦ C/min, a cooling to 25 ◦ C at the same rate, and a second heating equal
to the first one. One sample was tested for each composition and processing condition. The
glass transition temperature (Tg) of PET and the melting point (Tm) of both HDPE and PET
were determined during the first heating step, while the crystallization temperature of PET
(Tc) was determined during the cooling step. The degree of crystallinity of PET (Xc) was
calculated from the first heating cycle as expressed in Equation (1) [23]:

∆H f − ∆Hcc
Xc (%) = × 100% (1)
∆H 0f

where ∆H f is the enthalpy of fusion, ∆Hcc is the enthalpy of cold crystallization, and ∆H 0f
is the reference heat of melting of 100% crystalline PET (140.1 J/g [24]). The degree of
crystallinity of PET was also calculated during cooling and the second heating without any
cold crystallization component.
Scanning electron microscopy (SEM) observations of fractured surfaces were per-
formed with a Gemini-SEM 300 from Zeiss (Germany). The various specimens were dipped
in liquid nitrogen for about 15 min, then fractured in order to obtain observable surfaces.
A 20 nm gold layer was coated on all samples with a Quorum multi-coater. Observations
were carried out on filaments, printed, and injected specimens.

3. Results and Discussion

3.1. Thermal Stability
The TGA data are shown in Figure 1. Figure 1a reveals the absence of significant
amounts of water in the polymers, with no detectable evaporation around 100 ◦ C in spite
of the fact that the samples were not dried. The equilibrium water concentrations in HDPE
and PET (with a crystallinity of 52%) at 23 ◦ C and 50%RH are equal to 6 10−5 g/g [25] and
10−3 g/g [26], respectively. These amounts are below the detection limit of the TGA (HDPE)
and barely detectable (PET). The figure also confirms that no significant degradation of
7 5 of 14

07 processing
Polymers temperatures remained adapted for contaminated blends, as HDPE did5 not
2022, 14, 5507 of 145 of 15
degrade at 265 °C, even after 30 min.

the recycled polymers occurred below 300 ◦ C. Moreover, Figure 1b shows that the selected
processing temperatures remained
processing adapted
temperatures remainedfor adapted
contaminated blends, blends,
for contaminated as HDPE did not
as HDPE did not
degrade at 265 °C,degrade
even after 30
◦ min.
at 265 C, even after 30 min.

Figure 1. TGA and its derivatives of extruded PET and HDPE at 10 °C/min (a) and at 265 °C (b).

3.2. Quality of Extrusion and Printing

All 1.types
Figure TGAof filaments
and its were of
derivatives
Figure
extruded
1. TGA and extruded
its
and
PET
derivatives
printed
and HDPE
of extruded
withand
PETat
the same
10HDPE
°C/min process
◦ C/min
at 10(a) and atparameters.
(a)265
and°C (b).◦ C (b).
at 265
Interestingly, it was observed that the blends with the highest levels of HDPE contamina-
3.2. Quality of Extrusion and Printing
tion
3.2.(5Quality
and 10%) were
of Extrusioneasier to handle compared with filaments with lower amounts or
and Printing
without HDPE. Indeed,All thetypes
latterof filaments were extruded and printed with the same process parameters.
filaments often broke during winding, making it more
All types of filaments were
Interestingly, extruded
it was observed and printed
that the blends with thehighest
with the samelevels
process parameters.
of HDPE contamination
difficult to have a continuous
(5 and 10%) and
were easy
easier to
to roll
handlefilament.
compared Nonetheless,
with filaments all
with filaments,
lower regard-
amounts or without
Interestingly, it was observed that the blends with the highest levels of HDPE contamina-
less of the amount ofHDPE.
HDPE, could belatter
used in the printer andduring
produce high-quality prints.
tion (5 and 10%) were easier to handle compared with filaments with lower amountsdifficult
Indeed, the filaments often broke winding, making it more
or
The average diameter to have a continuous
and standard and easy for
deviation to roll filament.filaments
extruded Nonetheless, withall different
filaments, regardless
levels of
without HDPE. Indeed, the latter
the amount of HDPE, filaments
could beoften
used broke
in thethe during
printer andwinding,
produce making it prints.
high-quality more
of HDPE contamination are reported in Table 1, while filaments are shown in Figure The
difficult to have a continuous
average diameter and andeasy to roll deviation
standard filament.for Nonetheless, all filaments,
extruded filaments regard-
with different levels of
2. The 3D benchy models printed with these four Table
types1,of filaments are also shown in
less of the amount HDPE
of HDPE, could beare
contamination used in theinprinter
reported and
while produce high-quality
the filaments are shownprints.
in Figure 2.
Figure 2. All came out The 3Dwithout
benchy any
models trouble, withthese
printed with a quality that
four types was notare
of filaments affected
also shownby inthe
Figure 2.
The average diameter and standard deviation for extruded filaments with different levels
contamination. From Allacame
totaloutofwithout
13 dimensions,
any trouble,which were measured,
with a quality the average
that was not affected stand-
by the contamination.
of HDPE contaminationFrom
are reported
a total
in Tablewhich
of 13 dimensions,
1, while the filaments
were measured,
are shown in Figure
ard deviation between the measured and expected values was 1.6%the foraverage
both the standard
pure deviation
re-
2. The 3D benchy between
modelsthe printed
measured with
andthese
expected four types
values wasof1.6%
filaments
for both thearepure
alsorecycled
shownPET in and
cycled PET and the blend containing 2% of HDPE, 2.5% for the 5% HDPE blend, and 2.2%
Figure 2. All camethe out without
blend any2%
containing trouble,
of HDPE, with
2.5%aforquality that was
the 5% HDPE blend,notandaffected by one
2.2% for the thewith
for the one with 10%10% HDPE.
HDPE. Those
Those results
results were
were inthe
in thesame
same range
range as as the
the ones ones obtained
obtained for for
commercial
contamination. From a total of 13 dimensions, which were measured, the average stand-
commercial virgin (1.9%) and recycled
virgin (1.9%) and recycled (2.2%)(2.2%)filaments.
filaments.
ard deviation between the measured and expected values was 1.6% for both the pure re-
cycled PET and the blend containing 2% of HDPE, 2.5% for the 5% HDPE blend, and 2.2%
for the one with 10% HDPE. Those results were in the same range as the ones obtained for
commercial virgin (1.9%) and recycled (2.2%) filaments.

Figure 2. Filaments E-PET0,

Figure 2.E-PET2,
FilamentsE-PET5, and E-PET10,
E-PET0, E-PET2, fromE-PET10,
E-PET5, and left to from
rightleft
(a),
to and
right 3D benchy
(a), and 3D benchy
models printed with recycled PET contaminated
models printed withcontaminated
with recycled PET 0%, 2%, 5%,with
and0%,
10%2%,HDPE,
5%, andfrom10% left
HDPE,to right
from left to
(b). right (b).

Figure
3.3. 2. Filaments
Mechanical E-PET0, E-PET2, E-PET5, and E-PET10, from left to right (a), and 3D benchy
Properties
models printed with recycled PET contaminated with 0%, 2%, 5%, and 10% HDPE, from left to right
(b). The tensile behavior of all printed and injection-molded specimens is shown in Fig-
ure 3, and their Young’s modulus and maximum stress are reported in Table 3. The brittle
behavior of P-PET0
3.3. Mechanical is evident, with a strain at failure of 2.5%. The addition of 2%wt of
Properties
HDPE led to a slight increase in the Young’s modulus, but the samples remained brittle.
Polymers 2022, 14, 5507 6 of 15

3.3. Mechanical Properties

The tensile behavior of all printed and injection-molded specimens is shown in
Polymers 2022, 14, 5507 6 of 14
Figure 3, and their Young’s modulus and maximum stress are reported in Table 3. The
brittle behavior of P-PET0 is evident, with a strain at failure of 2.5%. The addition of 2 wt%
of HDPE led to a slight increase in the Young’s modulus, but the samples remained brittle.
failure, well beyond
In contrast, theofyield
additions 5 wt%point.
andSuch results
10 wt% led were unexpected increase
to a spectacular and contradict
in the the duc-
strain at
failure, well transition
tile-to-brittle beyond the yield point. observed
systematically Such results
withwere unexpected
increasing HDPE and contradict
fraction the
in the case
ductile-to-brittle
of transitionblends
immiscible HDPE-PET systematically observed
[12,13,27]. with increasing
These samples exhibitedHDPE fraction
a large drop in in the
the
case ofatimmiscible
stress HDPE-PET
2.5% strain, which then blends [12,13,27].
gradually These samples
decreased. exhibited
The reason a large
for this drop in
unexpected
the stress at mechanism
toughening 2.5% strain,inwhich thenofgradually
the case immiscibledecreased.
blends is The reason
further for this
explored inunexpected
the follow-
toughening mechanism in the case of immiscible blends is further explored in the following.
ing.

70 70
a) P-PET_C b) P-PET_C (3 prints)
60 60
P-PET0 P-PET0 (3 prints)
50 P-PET2 50 P-PET2 (3 prints)
Stress [MPa]

Stress [MPa]
40 P-PET5 40 P-PET5 (3 prints)
P-PET10 P-PET10 (3 prints)
30 30

20 20

10 10

0 0
0 5 10 15 20 0 5 10 15 20
Strain [%] Strain [%]
70 70
c) P-PET_CR_1 d) I-PET0
60 60 I-PET5
P-PET_CR_2
50 P-PET_CR_3 50
Stress [MPa]
Stress [MPa]

40 P-PET_CR_1 (3 prints) 40
P-PET-CR_2 (3 prints)
30 30
P-PET_CR_3 (3 prints)
20 20

10 10

0 0
0 5 10 15 20 0 5 10 15 20
Strain [%] Strain [%]

Figure
Figure 3. Stress-strain
Stress-strain curves
curves of
of (a)
(a) single
single prints,
prints, (b)
(b) three
three prints,
prints, (c) prints
prints of commercial
commercial recycled
recycled
filament (single prints and three prints), and (d) injected specimens.
filament (single prints and three prints), and (d) injected specimens.

TableThe Young’s
3. Young’s modulus
modulus and of the single-print
maximum blends
tensile stress was equal
for printed to 2.1 GPa for HDPE-PET
and injection-molded all HDPE
concentrations,
blends. higher than that of pure PET, which was equal to 1.7 GPa. The maximum
stress in these materials also increased with HDPE concentrations up to 5%wt, from about
45 MPa for P-PET0 to about 50 MPa. These values were lower
Single Prints Threethan those of specimens
Prints
Name printed with
Young’s Modulus the commercial, virgin
Max Stress PET filament and comparable
Young’s Modulus to the values
Max obtained
Stress
with[GPa]
the commercial, recycled [MPa] PET filament. However,
[GPa] these commercial filaments
[MPa] sys-
tematically led to brittle specimens, and only filaments containing HDPE led to more duc-
P-PET0 1.7 ± 0.3 44.8 ± 7.7 2.1 ± 0.1 47.9 ± 3.2
tile, single-printed samples.
P-PET2 2.1 ± 0.1 51.7 ± 1.1 2.2 ± 0.1 ± 0.9
Nevertheless, it turned out that printing artefacts had a major impact.48.5 Indeed, Figure
P-PET5 3b2.1 ± 0.0that the three-print
shows ± 1.7
49.5 samples 1.8 ± 0.2 printed simultaneously
(i.e., three specimens 37.1 ± 2.6 rather
P-PET10 than
2.1 one
± 0.2by one) were all45.6brittle,
± 1.8regardless of the2.1 amount
± 0.1 of HDPE used. Moreover,
34.4 ± 3.3 the
P-PET_C maximal
2.7 ± 0.1stress reached for P-PET5
65.1 ± 2.4 and P-PET10 was much
2.7 ± 0.1 lower than for P-PET0.
59.5 ± 2.1 The
number of simultaneously printed specimens was the only parameter that changed com-
P-PET_CR 2.0 ± 0.0 32.3 ± 13.7 1.9 ± 0.1 48.8 ± 2.3
pared to the samples analyzed previously. A possible explanation for the lack of ductility
I-PET0 in 2.3 ± 0.5
these samples is that the44.4 ± 4.2 layers cooled down to a lower temperature when mul-
printed
I-PET5 tiple
2.9 specimens
± 0.7 were printed34.2 simultaneously.
± 1.6 Indeed, printing a larger surface at the same
speed took more time, which means that the successive layers were applied to a cooler
surface. This difference in temperature presumably weakened the interface between two
printed layers, inducing a brittle fracture, as previously analyzed by Bakir et al. [28]. This
observation highlights the importance of adapted printing parameters and the considera-
ble sensitivity of 3D printing with recycled and contaminated materials. However, the
number of specimens that were simultaneously printed did not affect the properties of
samples printed with the commercial virgin PET filament.
Polymers 2022, 14, 5507 7 of 15

The Young’s modulus of the single-print blends was equal to 2.1 GPa for all HDPE
concentrations, higher than that of pure PET, which was equal to 1.7 GPa. The maximum
stress in these materials also increased with HDPE concentrations up to 5 wt%, from about
45 MPa for P-PET0 to about 50 MPa. These values were lower than those of specimens
printed with the commercial, virgin PET filament and comparable to the values obtained
with the commercial, recycled PET filament. However, these commercial filaments system-
atically led to brittle specimens, and only filaments containing HDPE led to more ductile,
single-printed samples.
Nevertheless, it turned out that printing artefacts had a major impact. Indeed,
Figure 3b shows that the three-print samples (i.e., three specimens printed simultane-
ously rather than one by one) were all brittle, regardless of the amount of HDPE used.
Moreover, the maximal stress reached for P-PET5 and P-PET10 was much lower than
for P-PET0. The number of simultaneously printed specimens was the only parameter
that changed compared to the samples analyzed previously. A possible explanation for
the lack of ductility in these samples is that the printed layers cooled down to a lower
temperature when multiple specimens were printed simultaneously. Indeed, printing a
larger surface at the same speed took more time, which means that the successive layers
were applied to a cooler surface. This difference in temperature presumably weakened the
interface between two printed layers, inducing a brittle fracture, as previously analyzed by
Bakir et al. [28]. This observation highlights the importance of adapted printing parameters
and the considerable sensitivity of 3D printing with recycled and contaminated materials.
However, the number of specimens that were simultaneously printed did not affect the
properties of samples printed with the commercial virgin PET filament.
Figure 3c shows the stress-strain curves of all single-print and three-print specimens
based on the commercial recycled filament. The first observation is the low repeatability of
the results. Indeed, two of the three single-print samples were brittle, as was one of the
three-print samples. The other three samples showed a more ductile behavior and up to
10% plastic deformation. The maximum stress reached, however, was much lower than for
virgin PET, with an average drop of about 15 MPa. The elastic modulus also significantly
decreased compared to virgin PET, dropping from 2.7 to about 2 GPa for both single-print
and three-print specimens. Even though even higher elongations at break were reached
in other studies [28], the values obtained here with a commercial recycled filament are
comparable to the behavior observed for homemade, contaminated recycled PET filaments.
Figure 3d shows that the injection-molded samples of pure recycled PET and recycled
PET with 5 wt% HDPE contamination were also brittle. The tensile strength of the pure
PET samples was 45 MPa, which corresponded to the value of similar materials [29]. The
addition of 5 wt% of HDPE led to a large reduction in tensile strength, in contrast to the
single-print samples.
To summarize, these results indicate that contamination with HDPE may lead to
a considerable and unexpected brittle-to-ductile transition in the otherwise brittle PET,
but only under specific printing conditions. The contamination reduced the strength of
injection-molded samples, leading to properties that are similar to those of printed recycled
PET in the 0◦ direction with 100% infill.

3.4. Microstructural Analyses

HDPE is immiscible with PET, as clearly observed on the micrographs of extruded
filaments in Figure 4. There are spherical HDPE inclusions with diameters ranging from a
few hundred nanometers to two micrometers that are not bonded to the PET. Inclusions
smaller than 1 µm were present in the filaments E-PET2 and E-PET5, whereas inclusions in
E-PET10 were all larger than 3 µm. These morphologies are characteristic of immiscible
polymer blends, as reported in numerous works [12,13,30,31].
 material [12,13,31]. In printed specimens, small inclusions found in small amounts, like in
the 2%wt HDPE contaminated sample, weakened the material as well. However, when
the number of these inclusions increased (P-PET5 and P-PET10), HDPE acted locally as a
plasticizer and allowed the material to undergo plastic deformation. As the microstruc-
ture14,did
Polymers 2022, 5507not change between single-print and three-print samples, the brittleness observed
8 of 15
in the latter case could be attributed to poor interlayer adhesion.

Figure 4. Electron
Figure 4. Electron micrographs micrographs
of extruded of extruded
filaments filaments
for (a) E-PET0,for(b)
(a) E-PET2,
E-PET0, (b)
(c)E-PET2,
E-PET5,(c)and
E-PET5,
(d) and
E-PET10. (d) E-PET10.

Quite remarkably, the HDPE inclusions were elongated and thinner in all printed
specimens, as shown in Figure 5. This unique morphology is likely to result from the shear
forces induced during printing within a die smaller than the filament diameter, causing the
observed deformation of the HDPE inclusions [13]. The same tendency was observed for
both single-print and three-print batches, with similar elongated inclusions for both P-PET5
specimens (Figure 5b,d). Moreover, the size of the inclusions, around a few hundreds of
nanometers, did not seem to depend on the contamination level, whereas their number
increased accordingly. This is because the droplet size mostly depends on shear rate,
interfacial tension, and the viscosity of the two phases and not on concentration [30]. Aside
from this, a gap was observed between the PET and HDPE phases for all types of printed
specimens. This could be attributed to a difference in crystallization shrinkage and thermal
contraction between these two polymers during the fast cooling, as this phenomenon was
not observed in the filament or the injected specimens, which had a slower cooling rate. In
the latter case, large HDPE inclusions in the range 1–4 µm were observed in fewer numbers
than in printed specimens, as seen in Figure 6b for the sample contaminated with 5 wt% of
HDPE, presumably owing to different shear rates between the two types of processes [30].
The morphology of the PET phase shown in Figure 6 was comparable between the pure
specimen (I-PET0) and the contaminated one (I-PET5).
4, 5507 Polymers 2022, 14, 5507 9 of 14 9 of 15

Figure 5. Electron
Figure 5. Electron micrographs micrographs
of single of P-PET0,
prints (a) single prints (a) P-PET0,
(b) P-PET5 and(b)three
P-PET5 and (c)
prints, three prints, (c) P-PET0,
P-PET0,
(d) P-PET2,
(d) P-PET2, (e) P-PET5, and (f)(e) P-PET5, and (f) P-PET10.
P-PET10.

Figure 6. Electron
Figure 6. Electron micrographs micrographs
of injected of injected
specimens specimens
(a) I-PET0 and (a)
(b)I-PET0 and (b) I-PET5.
I-PET5.

The peculiar microstructures could explain the mechanical behavior of contaminated

3.5. Thermal Transitions and Crystallinity
recycled PET. The large HDPE inclusions found in the injected specimen weakened the
The crystallinity in the various blends was investigated using DSC to better under-
stand the interplay between the observed morphologies and the brittle-to-ductile transi-
tion. Figure 7 shows the thermogram of a part printed with a filament of PET contami-
nated with 5%wt HDPE (P-PET5), where the main transitions of the two phases are la-
 Polymers 2022, 14, 5507 10 of 15

material [12,13,31]. In printed specimens, small inclusions found in small amounts, like in
the 2 wt% HDPE contaminated sample, weakened the material as well. However, when
the number of these inclusions increased (P-PET5 and P-PET10), HDPE acted locally as a
plasticizer and allowed the material to undergo plastic deformation. As the microstructure
did not change between single-print and three-print samples, the brittleness observed in
the latter case could be attributed to poor interlayer adhesion.

3.5. Thermal Transitions and Crystallinity

The crystallinity in the various blends was investigated using DSC to better understand
Polymers 2022, 14, 5507 the interplay between the observed morphologies and the brittle-to-ductile transition.
Figure 7 shows the thermogram of a part printed with a filament of PET contaminated with
5 wt% HDPE (P-PET5), where the main transitions of the two phases are labeled.

12
Crystallization
9 PET

Crystallization
6 HDPE

3
Heat flow [J/g]

Cold-crystallization
0 PET

-3
Glass
transition Melting
-6 PET HDPE

-9
Melting PET
-12
0 40 80 120 160 200 240 280 320
Temperature [ºC]

Heating
Figure 7.7.
Figure (red (red
Heating curve,curve,
down) and cooling
down) and(blue curve, up)
cooling thermograms
(blue of athermograms
curve, up) part printed withof a part
a filament of PET contaminated with 5 wt% HDPE.
with a filament of PET contaminated with 5%wt HDPE.
The first heating thermograms of all investigated materials are reproduced in Figure 8,
The first heating
and corresponding thermalthermograms
transitions andofPET all crystallinity
investigated arematerials
summarized are
inreproduced
Table 4. in
The degree of crystallinity of the pure, non-contaminated PET was
8, and corresponding thermal transitions and PET crystallinity are summarized iquite different between
the different processes, with the single-print samples with rapid cooling (P-PET0 1×) being
4. The degree of crystallinity of the pure, non-contaminated PET was quite diffe
much lower (13%) than the other three types of samples, which experienced slower cooling
tween
(around the different processes, with the single-print samples with rapid cooling (
20%).
1x) being
Regardingmuch lower (13%)specimens,
the contaminated than the the otherfact three
that thetypes
meltingof point
samples,
of HDPEwhich
is expe
slower cooling (around 20%).
close to the cold crystallization of PET makes the interpretation of the thermograms more
challenging, especially
Regarding thewhen the amount of
contaminated HDPE becomes
specimens, thesignificant
fact that(above 5%). Indeed,
the melting point of H
both transitions occurred around 120 ◦ C, which must be considered when interpreting
close to the cold crystallization of PET makes the interpretation of the thermogram
the values reported in Table 4. In general, the degree of crystallinity appears to increase
challenging,
with increasingespecially when the
HDPE concentration amount
in both filamentsof HDPE becomes
and printed significant
parts. Indeed, HDPE(above 5
deed, both
inclusions maytransitions
have acted occurred around
like germination 120for
points °C,crystallization.
which mustOnly be considered
the results forwhen in
the injected parts were unexpected, as the degree of crystallinity of I-PET5
ing the values reported in Table 4. In general, the degree of crystallinity appear was one of the
lowest observed in this study, namely 16%. The crystallinity of the PET itself (calculated
crease with increasing HDPE concentration in both filaments and printed parts.
from the cooling curve) was constant at around 26%, no matter the process or the amount
HDPE inclusions
of contamination. may
This valuehave acted like
is comparable germination
to recently reported points for crystallization.
crystallinities on filaments Only
sults for the
made from injected
extruded parts [14,32].
PET bottles were unexpected, as the degree
Only injected specimens of acrystallinity
displayed larger degree of I-PE
of crystallinity (around 35%) for both I-PET0 and I-PET5.
one of the lowest observed in this study, namely 16%. The crystallinity of the PE
(calculated from the cooling curve) was constant at around 26%, no matter the pro
the amount of contamination. This value is comparable to recently reported crysta
on filaments made from extruded PET bottles [14,32]. Only injected specimens di
a larger degree of crystallinity (around 35%) for both I-PET0 and I-PET5.
In summary, neither the process nor the HDPE contamination significantly a
the glass and melting temperatures of the recycled PET. Knowing this and cons
 38]. This interval was more than three times longer when three samples were printed sim-
ultaneously. The interface temperature between adjacent layers at the time of printing was
therefore higher for single prints, which favored interlayer adhesion and reduced the in-
terlayer stress [39–41]. If poor interlayer adhesion was the reason for the brittleness of
some
Polymers 2022, printed recycled PET samples, crystallinity was not involved, as confirmed by the
14, 5507 11 of 15
performance of the commercial PET samples, which were not crystalline.

Figure 8. First heating thermograms of extruded, printed (single prints 1× and three prints 3×), and
Figure 8. First heating thermograms
injection molded PET of extruded, printed
filaments and (single
samples prints 1x
with different andof
levels three
HDPEprints 3x), and and of
contamination
injection molded PET filaments
virgin and commercial
and recycled samples withPET,different levels
both extruded andofprinted.
HDPE contamination and of
virgin and recycled commercial PET, both extruded and printed.
Table 4. Glass transition Tg, and melting point Tm, of HDPE and PET and crystallinity of PET in
Indeed, as shown
blends.in Figure
Single 8,are
prints notindicated
only theas ‘1commercial virgin
×’ and three prints as ‘3and recycled filaments
×’, respectively.
but also the parts printed with these two materials were completely amorphous. This is a
Tg PET Tm HDPE Tm PET Xc PET
major difference(◦between
Name
C)
homemade (◦ C)
and commercial (◦ C)
filaments and prints. As all printed
1st Heating Cooling
specimens
E-PET0
were processed
73.4
with the-
exact same parameters,
253.1
it suggests 19%
that the crystallin-
27%
ity of the commercial
E-PET2 75.6 filaments was related to their252.3
128.3 composition rather 20%than to processing 27%
conditions (e.g., melting temperature or cooling rate). The only consequence of processing
E-PET5 73.6 128.1 251.7 25% 25%
E-PET10 74.0 129.1 252.2 23% 26%
seemed to be the lowering of the glass transition temperature of the virgin PET filament
I-PET0 74.2 -
when it was printed (going from 76.1 to 68.4 °C), 255.8
as shown in Table 23%
4. As mentioned 35%
in
I-PET5 75.1 129.4 256.4 16% 32%
the introduction, one of the main challenges of using PET as a feedstock for 3D printing is
P-PET0 (1×) 76.4 - 251.6 13% 25%
P-PET2its(1crystallinity,
×) so it is not surprising
74.8 No signalthat commercial 252.3 filaments are19% developed in a way 25%
P-PET5that
(1×avoids
) crystallization.
74.5 Nevertheless,
128.6 the fact that252.4similar-quality23% devices could be26% ob-
P-PET10 (1 × ) 73.9 128.5 252.3
tained with recycled PET contaminated with significant amounts of HDPE, regardless of 22% 27%
P-PET0 ×)
its(3crystallinity, 75.7 a promising observation
was - 252.8 developments.
for further 18% 29%
P-PET2 (3×) 73.9 No signal 252.8 18% 27%
P-PET5 (3×) 76.3 128.1 252.7 21% 26%
Table
P-PET10 (3×) 4. Glass transition
75.9 Tg, and melting
129.3 point Tm, of 253.7
HDPE and PET and22%
crystallinity of PET in
27%
blends. Single prints
PET_C 76.1
are indicated as ‘1x’
-
and three prints-as ‘3x’, respectively.
0% 0%
PET_CR 78.1 - - 0% 0%
P-PET_C (1×)Name 68.4Tg PET Tm HDPE
- Tm PET- Xc
0% PET 0%
P-PET_CR (1×) 79.2 (°C) (°C)
- (°C) - 1st Heating0% Cooling 0%
P-PET_C (3×) 68.1 - - 0% 0%
E-PET0
P-PET_CR (3×) 78.4
73.4 -- 253.1 - 19% 0% 27% 0%
E-PET2 75.6 128.3 252.3 20% 27%
E-PET5 73.6 128.1 251.7 25% 25%
Polymers 2022, 14, 5507 12 of 15

In summary, neither the process nor the HDPE contamination significantly affected
the glass and melting temperatures of the recycled PET. Knowing this and considering
that the amount of HDPE was not high enough to influence the melting point of PET, the
printing and injection temperatures did not have to be adapted when HDPE was added to
the recycled PET blends.
A correlation was observed between crystallinity and the mechanical properties, as
reported in previous investigations [14,33–35]. Indeed, the printed samples that were able
to undergo plastic deformation (P-PET5 and P-PET10 single prints) were those with the
highest degree of crystallinity. However, this tendency was not observed for the three prints
and the injected samples, which were brittle regardless of crystallinity and contamination.
The fact that no significant differences in the degree of crystallinity were observed between
single-print and three-print samples indicates that their brittleness was not only related
to crystallinity. Even though layers were printed on a cooler surface when three samples
were printed simultaneously, the degree of crystallinity of recycled PET contaminated
with HDPE did not change. Indeed, there were 27 s of delay between the printing of two
consecutive layers with single printing, which corresponds to the time for the filament to
cool down from the die temperature (265 ◦ C) to the bed temperature (100 ◦ C) [36–38]. This
interval was more than three times longer when three samples were printed simultaneously.
The interface temperature between adjacent layers at the time of printing was therefore
higher for single prints, which favored interlayer adhesion and reduced the interlayer
stress [39–41]. If poor interlayer adhesion was the reason for the brittleness of some printed
recycled PET samples, crystallinity was not involved, as confirmed by the performance of
the commercial PET samples, which were not crystalline.
Indeed, as shown in Figure 8, not only the commercial virgin and recycled filaments
but also the parts printed with these two materials were completely amorphous. This is a
major difference between homemade and commercial filaments and prints. As all printed
specimens were processed with the exact same parameters, it suggests that the crystallinity
of the commercial filaments was related to their composition rather than to processing
conditions (e.g., melting temperature or cooling rate). The only consequence of processing
seemed to be the lowering of the glass transition temperature of the virgin PET filament
when it was printed (going from 76.1 to 68.4 ◦ C), as shown in Table 4. As mentioned in
the introduction, one of the main challenges of using PET as a feedstock for 3D printing is
its crystallinity, so it is not surprising that commercial filaments are developed in a way
that avoids crystallization. Nevertheless, the fact that similar-quality devices could be
obtained with recycled PET contaminated with significant amounts of HDPE, regardless of
its crystallinity, was a promising observation for further developments.

4. Conclusions
The present investigation of the influence of HDPE contamination and process param-
eters on the microstructure, mechanical performance, and printing quality of 3D printed
parts made of recycled bottle-grade PET leads to the following conclusions.
The contamination with HDPE, which came mainly from rings and caps, which are
available in the same waste stream, did not affect the ease of extrusion or the printing quality.
At contamination levels below 2 wt%, the brittleness of the printed samples increased,
with a strain at failure of 2.5%, but these were highly ductile at higher contamination
levels, with a strain at failure above 15%. Moreover, the elastic modulus of PET with
5 wt% of HDPE was higher than that of pure recycled PET. Such considerable toughening
behavior was attributed to the presence of elongated HDPE inclusions in the immiscible
blend resulting from shear forces induced by the printing, and to differences in the PET
crystalline structure. This was in fact true only for specific temperature conditions during
the printing of successive layers. In the case of three parts printed in parallel with an 85 s
cooling time between adjacent layers, the lower temperature of the interface resulted in
weaker interfaces and brittle parts. In contrast, when single parts were printed sequentially
Polymers 2022, 14, 5507 13 of 15

with a cooling time of 27 s, high enough interfacial temperatures could be maintained

throughout the printing cycle.
This investigation confirmed that the mechanical performance of recycled PET was
very sensitive to the processing conditions, especially in the case of 3D printing. However,
a remarkable toughening effect was uncovered in the contaminated blend, in spite of
its immiscible character, due to a process-induced microscale morphology and favorable
interface temperature conditions between adjacent layers. This demonstrated that a PET
recycling process that was tolerant of HDPE contamination and avoided sophisticated
compatibilization schemes was possible. With the processing limitations requiring adapted
printing strategies, such a low-cost process enabled the printing of quality parts with
accuracies comparable to those which can be obtained with high purity, commercially
recycled filaments. In future work, both practical and fundamental issues should be
considered. Firstly, the recycling logistics, which include waste collection and sorting, must
be addressed to ensure that the contamination level remains acceptable so that adequate
3D printing applications can be identified. Secondly, a more systematic investigation of
the influence of the process conditions should be conducted. The focus should be on the
speed of printing, the resulting shape of the HDPE nodules, and the interface temperature.
This information would enable us to better understand the key process influences on the
mechanical performance of complex 3D-printed parts.

Author Contributions: Conceptualization J.V., A.D., V.M. and Y.L.; experimental investigation J.V.
and A.D.; writing—original draft preparation J.V.; writing—review and editing, J.V., A.D., V.M. and
Y.L.; supervision, project administration, and funding acquisition, V.M. and Y.L. All authors have
read and agreed to the published version of the manuscript.
Funding: This work was funded by the EPFL TECH4DEV fund 2020: SUSTAINABLE 3D PRINTING
FOR PEACE IN COLOMBIA—3DP4PEACE.
Institutional Review Board Statement: Not applicable.
Data Availability Statement: All experimental data are available upon reasonable request to the
corresponding author.
Acknowledgments: We would like to thank the following LPAC colleagues for help with experiments:
Nour Halawani (SEM), Raymond de Ribains (SEM), Aigoul Schreier (SEM preparation, TGA), and
Alexandre A. Looten (TGA).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. PlasticsEurope. Plastics—The Facts 2020: An Analysis of European Plastics Production, Demand and Waste Data. Available online:
https://plasticseurope.org/wp-content/uploads/2021/09/Plastics_the_facts-WEB-2020_versionJun21_final.pdf (accessed on 3
August 2022).
2. World Economic Forum. The New Plastics Economy: Rethinking the Future of Plastics. Available online: https://www3
.weforum.org/docs/WEF_The_New_Plastics_Economy.pdf (accessed on 3 August 2022).
3. Geyer, R.; Jambeck, J.R.; Law, K.L. Production, use, and fate of all plastics ever made. Sci. Adv. 2017, 3, e1700782. [CrossRef]
[PubMed]
4. Welle, F. Twenty years of PET bottle to bottle recycling—An overview. Resour. Conserv. Recycl. 2011, 55, 865–875. [CrossRef]
5. Benyathiar, P.; Kumar, P.; Carpenter, G.; Brace, J.; Mishra, D.K. Polyethylene Terephthalate (PET) Bottle-to-Bottle Recycling for the
Beverage Industry: A Review. Polymers 2022, 14, 2366. [CrossRef] [PubMed]
6. Thakur, N.; Rajput, K.; Kaur, N.; Mehta, P. A Persistence Search for the Most Appropriate Process of PET Recycling: A Review. J.
Pharm. Res. Int. 2021, 33, 681–688. [CrossRef]
7. Damayanti; Wu, H.-S. Strategic Possibility Routes of Recycled PET. Polymers 2021, 13, 1475. [CrossRef]
8. R-PET: Schweizer Kreislauf—PET-Recycling. Available online: https://www.petrecycling.ch/fr/savoir/recycling-pet/r-pet-
schweizer-kreislauf-kopie (accessed on 1 July 2022).
9. Pinter, E.; Welle, F.; Mayrhofer, E.; Pechhacker, A.; Motloch, L.; Lahme, V.; Grant, A.; Tacker, M. Circularity Study on PET
Bottle-To-Bottle Recycling. Sustainability 2021, 13, 7370. [CrossRef]
10. Navarro, R.; Ferrándiz, S.; López, J.; Seguí, V.J. The influence of polyethylene in the mechanical recycling of polyethylene
terephtalate. J. Mater. Process. Technol. 2008, 195, 110–116. [CrossRef]
Polymers 2022, 14, 5507 14 of 15

11. Todd, A.D.; McEneany, R.J.; Topolkaraev, V.A.; Macosko, C.W.; Hillmyer, M.A. Reactive Compatibilization of Poly(ethylene
terephthalate) and High-Density Polyethylene Using Amino-Telechelic Polyethylene. Macromolecules 2016, 49, 8988–8994.
[CrossRef]
12. Pracella, M.; Rolla, L.; Chionna, D.; Galeski, A. Compatibilization and properties of poly(ethylene terephthalate)/polyethylene
blends based on recycled materials. Macromol. Chem. Phys. 2002, 203, 1473–1485. [CrossRef]
13. Kratofil Krehula, L.; Hrnjak-Murgić, Z.; Jelenčić, J. Study of masterbatch effect on miscibility and morphology in PET/HDPE
blends. J. Adhes. Sci. Technol. 2015, 29, 74–93. [CrossRef]
14. Van de Voorde, B.; Katalagarianakis, A.; Huysman, S.; Toncheva, A.; Raquez, J.M.; Duretek, I.; Holzer, C.; Cardon, L.; Bernaerts,
K.V.; Van Hemelrijck, D.; et al. Effect of extrusion and fused filament fabrication processing parameters of recycled poly(ethylene
terephthalate) on the crystallinity and mechanical properties. Addit. Manuf. 2022, 50, 102518. [CrossRef]
15. Zander, N.E.; Gillan, M.; Lambeth, R.H. Recycled polyethylene terephthalate as a new FFF feedstock material. Addit. Manuf. 2018,
21, 174–182. [CrossRef]
16. Vaes, D.; Van Puyvelde, P. Semi-crystalline feedstock for filament-based 3D printing of polymers. Prog. Polym. Sci. 2021, 118,
101411. [CrossRef]
17. Tsai, H.-H.; Wu, S.-J.; Wu, Y.-D.; Hong, W.-Z. Feasibility Study on the Fused Filaments of Injection-Molding-Grade Poly(Ethylene
Terephthalate) for 3D Printing. Polymers 2022, 14, 2276. [CrossRef]
18. Schneevogt, H.; Stelzner, K.; Yilmaz, B.; Abali, B.E.; Klunker, A.; Völlmecke, C. Sustainability in additive manufacturing: Exploring
the mechanical potential of recycled PET filaments. Compos. Adv. Mater. 2021, 30, 26349833211000064. [CrossRef]
19. Prior, M. PLA vs. PETG: Which Material Should You Choose? 3Dnatives, 11 May 2021. Available online: https://www.3dnatives.
com/en/pla-vs-petg-which-material-should-you-choose-110520215/ (accessed on 4 August 2022).
20. La Mantia, F.P.; Vinci, M. Recycling poly(ethyleneterephthalate). Polym. Degrad. Stab. 1994, 45, 121–125. [CrossRef]
21. CreativeTools. #3DBenchy—The Tool to Calibrate and Test Your 3D Printer. Instructables. Available online: https://www.
instructables.com/3DBenchy-The-tool-to-calibrate-and-test-your-3D-pr/ (accessed on 1 July 2022).
22. ISO 527-1:2019; Plastics—Determination of Tensile Properties—Part 1: General Principles; Technical Committee ISO/TC 61/SC 2
Mechanical Behavior. International Organization for Standardization: Geneva, Switzerland, 2019.
23. Fakirov, S.; Banerjee, S.; Lin, R.J.T. On the Degree of Crystallinity from DSC in the Case of Multiple Melting of Synthetic Polymers.
J. Macromol. Sci. Part B 2007, 46, 317–320. [CrossRef]
24. Torres, N.; Robin, J.J.; Boutevin, B. Study of thermal and mechanical properties of virgin and recycled poly(ethylene terephthalate)
before and after injection molding. Eur. Polym. J. 2000, 36, 2075–2080. [CrossRef]
25. Suherman, S.; Peglow, M.; Tsotsas, E. Measurement and Modelling of Sorption Equilibrium Curve of Water on PA6, PP, HDPE
and PVC by Using Flory-Huggins Model. Reaktor 2010, 13, 89–94. [CrossRef]
26. Shigetomi, T.; Tsuzumi, H.; Toi, K.; Ito, T. Sorption and diffusion of water vapor in poly(ethylene terephthalate) film. J. Appl.
Polym. Sci. 2000, 76, 67–74. [CrossRef]
27. Iniguez, G.C.; Michel, E.; Gonzalez-Romero, V.M.; Gonzalez-Nunez, R. Morphological stability of postconsumer PET/HDPE
blends. Polym. Bull. 2000, 45, 295–302. [CrossRef]
28. Bakır, A.A.; Atik, R.; Özerinç, S. Effect of fused deposition modeling process parameters on the mechanical properties of recycled
polyethylene terephthalate parts. J. Appl. Polym. Sci. 2021, 138, 49709. [CrossRef]
29. Alsewailem, F.D.; Alrefaie, J.K. Effect of contaminants and processing regime on the mechanical properties and moldability of
postconsumer polyethylene terephthalate bottles. Waste Manag. 2018, 81, 88–93. [CrossRef]
30. Sundararaj, U.; Macosko, C.W. Drop Breakup and Coalescence in Polymer Blends: The Effects of Concentration and Compatibi-
lization. Macromolecules 1995, 28, 2647–2657. [CrossRef]
31. Wyser, Y.; Leterrier, Y.; Månson, J.-A.E. Effect of inclusions and blending on the mechanical performance of recycled multilayer
PP/PET/SiOx films. J. Appl. Polym. Sci. 2000, 78, 910–918. [CrossRef]
32. Bustos Seibert, M.; Mazzei Capote, G.A.; Gruber, M.; Volk, W.; Osswald, T.A. Manufacturing of a PET Filament from Recycled
Material for Material Extrusion (MEX). Recycling 2022, 7, 69. [CrossRef]
33. Viana, J.C.; Alves, N.M.; Mano, J.F. Morphology and mechanical properties of injection molded poly(ethylene terephthalate).
Polym. Eng. Sci. 2004, 44, 2174–2184. [CrossRef]
34. Woods, D. Effects of crystallization on the glass-rubber transition in polyethylene terephthalate filaments. Nature 1954, 174,
753–754. [CrossRef]
35. Ward, I.M. The mechanical behavior of poly(ethylene terephthalate). J. Macromol. Sci. Part B 1967, 1, 667–694. [CrossRef]
36. Dinwiddie, R.B.; Kunc, V.; Lindal, J.M.; Post, B.; Smith, R.J.; Love, L.; Duty, C.E. Infrared imaging of the polymer 3D-printing
process. Thermosense Therm. Infrared Appl. XXXVI 2014, 9105, 910502.
37. Zhou, Y.; Nyberg, T.; Xiong, G.; Liu, D. Temperature analysis in the fused deposition modeling process. In Proceedings of the
2016 3rd International Conference on Information Science and Control Engineering (ICISCE), Beijing, China, 8–10 July 2016; pp.
678–682.
38. El Moumen, A.; Tarfaoui, M.; Lafdi, K. Modelling of the temperature and residual stress fields during 3D printing of polymer
composites. Int. J. Adv. Manuf. Technol. 2019, 104, 1661–1676. [CrossRef]
39. Ferraris, E.; Zhang, J.; Van Hooreweder, B. Thermography based in-process monitoring of Fused Filament Fabrication of polymeric
parts. CIRP Ann. 2019, 68, 213–216. [CrossRef]
Polymers 2022, 14, 5507 15 of 15

40. Lepoivre, A.; Boyard, N.; Levy, A.; Sobotka, V. Heat transfer and adhesion study for the FFF additive manufacturing process.
Procedia Manuf. 2020, 47, 948–955. [CrossRef]
41. Vanaei, H.R.; Shirinbayan, M.; Deligant, M.; Khelladi, S.; Tcharkhtchi, A. In-process monitoring of temperature evolution during
fused filament fabrication: A journey from numerical to experimental approaches. Thermo 2021, 1, 332–360. [CrossRef]